# Gaussian Process regression: does there exist a conjugate prior over hyperparameters?

When adopting a fully Bayesian hierarchical setting in Gaussian Process regression is there a choice of kernel (covariance) function such that there exist a conjugate prior? If so which?